Data Analysis

Water quality samples and associated data were collected on June 7, 8, and 15th, 2008.   The UV index was 9, 8, and 8 on these days respectively.   General weather conditions were similar, but not identical, for each of the sampling events.   The weather was characterized by (a) bright sun in the morning followed by increasing clouds in the afternoon, (b) wind generally out of the southwest, and (c) above average air temperature for this time of year, resulting in many swimmers.  Detailed observations and measurements are provided in our formal report; This Blog entry presents a summary of our results.  

The attempt to use Solar Print Paper to quantify the ultraviolet radiation was not successful.   The paper goes to dark blue with any sun exposure i.e. it does not have shades of blue which could be used to quantify the amount of ultraviolet radiation.

Raw bacterial counts (#) were converted to concentration (# per 100 ml ) by the following formula:

     The volume of sample plated ranged from 2.61 +/- 0.03 ml on June 7th & 8th to 2.42 +/‑ 0.02 ml on June 15th.  The change in volume plated was due to a change in pipette type.  The volume of each pipette was measured by delivering an aliquot of distilled water onto a digital balance and recording mass.  The mass was converted to volume by assuming that the density of the distilled water was 1.0 g. per ml.  Results were measured in triplicate to assess repeatability of the volume delivered by each pipette.

     We chose to focus on fecal coliform levels rather than total coliform levels because we believe fecal coliform is a better indicator of water contamination.  As noted by Microbiology Labs, LLC.,  “Non-fecal coliforms are widely distributed in nature, being found both as naturally-occurring soil organisms and in the intestines of warm-blooded animals and humans…..Fecal coliform…is the result of some form of fecal contamination.  Sources may be either animal or human.”  

Statistical Significance of Variation

Three samples were taken at each location with one plate prepared for each sample.   The data in Table 1 is the bacterial count average for the three plates from each sample location.  

Table 1:  Summary of Mean Fecal Coliform Concentrations (#/100 ml)

 by sampling location and time

Taking multiple samples allowed us to estimate the variability involved in sampling and testing.    The pooled standard deviation across the eighteen unique combinations of sample day and sample location was 58.3 colonies/100 ml.    Based on a sample size of 3, a 95% confidence interval around the mean bacterial count is +/- 67.4 colonies/100 ml.   The nature and scope of this project did not allow us to separate the variability into sampling component and testing components.  Note that the only statistically significant differences between the a.m. and p.m. samples at a specific location were Locations 1 and 2 on Day 1 and Location 3 on Day 2.   The other differences between the a.m. and p.m. samples on a given day at a given location are within the sampling and testing variability.

Diurnal Variations in Water Quality

     The first element of our research question was “How does water quality as indicated by bacterial count vary over the course of the day?”   Contrary to our expectations based on our hypothesis about UV exposure, the data indicated that bacterial counts were generally higher at 2pm than at 9am.  Figure 1 presents the average of 9 bacterial level measurements for each sampling event as a summary of diurnal variation. 

 Figure 1

Correlations between bacterial concentration and other variables

The second element of our research question was “Can diurnal variations in bacteria concentration be correlated with other variables?”   In order to explore this question, we measured algae level, water current direction and velocity, water temperature, pH, conductivity, turbidity, and dissolved oxygen at each sample location.  Dissolved oxygen (DO) was measured on the first two sampling days, but was not recorded on the third sampling day to due anomalies in measured values.  We recorded overall UV index, cloud cover, air temperature, river plume direction, wind direction, and wind speed at the time of sampling.  General observations such as beach conditions and a qualitative assessment of the number of people using the beach were also made.   

Algae level and water current data and observations were analyzed for each sample location on each day.   The observations for each day are summarized and compared graphically to fecal coliform data in our formal report.  

Algae Level

Algae concentrations were observed at regular intervals from the shoreline to the point at which the samples were taken. The intervals included 0 to 3 feet, 3 to 6 feet, and 6 feet from the shoreline to the sampling location.  The sampling location was approximately 20 to 30 feet from shore, depending on the water depth.  Specific observations were recorded using descriptors such as, clear, “floating tufts”, cloudy, pea soup.  

             The algae level varied significantly between sampling locations.  The west end of the beach typically contained the lowest algae levels, and the east end of the beach typically contained the highest.   This was expected because of the location of the pier on the east end.  Algae levels also varied with distance from shore, with generally higher concentrations closer to the shore.  No significant correlation was observed between fecal coliform concentration and local algae level.

Water Current Velocity and Direction

Wind has the biggest impact on surface water currents.   Wind direction is recorded in degrees relative to true North.  The beach runs from NW to SE with the western end at approximately 315 degrees and the eastern end at 135 degrees relative to true north.  The wind direction ranged from 210 degrees to 270 degrees during the sampling periods for this study.   The wind speed is reported by the weather station in meters/second (m/s) and was corrected to miles/hour (mph) by multiplying raw data by 2.236.   The wind speeds ranged from 2.2 mph to 21.3 mph during sampling periods.

We observed that water current patterns were quite complicated.  Our protocol involved measuring water current and direction at each sampling location.  We collected a single point sample, but observed complex patterns around the sampling location.   On the morning of Day 1, the current at all three locations showed a general flow away from the beach but by the afternoon sampling the wind had shifted 20 degrees to the north and increased in speed which resulted in a general flow into the swimming area and fecal coliform levels showed a statistically significant increase at Locations 1 and 2.    Location 3 did not show a significant increase but the mean did increase slightly. 

On the morning of Day 2, the wind was almost directly perpendicular to the shoreline and from the beach.   Current observations indicated a general outflow from the swimming area.   By the afternoon, the wind had shifted 60 degrees toward true north resulting, once again, in a flow toward the eastern end of the beach and statistically higher fecal coliform levels in Location 3.

On Day 3, the wind shifted 10 degrees away from true north between the morning and afternoon sample, which prolonged the general outflow observed in the morning.  During the afternoon sampling period, the wind speed dropped considerably.  The differences at each location between morning and afternoon bacterial levels were not statistically significant on Day 3.

On Day 3 at Location 3, the wind direction was affected more by the river gorge than the prevailing winds.   We noted that the flag at both the Coast Guard Station and the Rochester Yacht Club were aligned parallel to the river;  a different position than of the flag near the beach house which reflected the prevailing wind.   We believe that wind direction affected by the river gorge affected the surface current at Location 3 and caused more outward flow from the area keeping bacteria levels low relative to other sampling days.

We also noted that waves bouncing back from the pier frequently were perpendicular to the incoming waves from the general surface current.    There were some areas along the beach where the waves combined to create relatively stagnant areas; at these locations, cross currents caused broken shells to build up at the edge of the water at regular intervals along the beach.

Improvements in measurement techniques along with more detailed observations may reveal a stronger correlation between bacterial level and water current velocity/direction.

Water Temperature 

Water temperature ranged from 13.0 to 19.8 deg C between June 7th and June 15th, following a generally upward trend.  No significant correlation between water temperature and bacterial concentration was observed.

 pH

pH ranged from 8.1 to 8.8 between June 7th and June 15th, following a generally upward trend.  No significant correlation between pH and bacterial concentration was observed.   

Conductivity

Conductivity ranged from 108 to 195 microsiemens  per centimeter, in what appeared to be a random pattern.  No significant correlation between conductivity and bacterial concentration was observed.   

Dissolved Oxygen

As described previously, difficulties were encountered in the dissolved oxygen (DO) calibration process.  We do not have a high level of confidence in the validity of our dissolved oxygen measurements.  For this reason we discontinued DO measurements after the second sample day, and suggest calibration procedure improvements as a follow-up to this research. 

UV index

The UV index ranged from a value of 9 on June 7th to a value of 8 on June 8th and 15th.  Unfortunately there was insufficient variation in the UV index to draw conclusions about the effect of UV index on water quality.  UV index can be considered a relatively consistent factor during data collection for this study. 

 Cloud Cover

The average cloud cover was 33% on June 7th, 56% on June 8th, and 53% on June 15th.  This variable was generally correlated with UV index, but a relationship between cloud cover and bacterial level could not be established.

Air Temperature

Air temperature ranged from 21.5 deg C on June 15th to 31.3 deg C on June 8th.  A relationship between air temperature and bacterial level could not be established.    

River Plume Direction

The river plume direction was toward the east on June 7th and 8th, and toward the west on June 15th.   A relationship between river plume direction and bacterial level could not be established.    

Beach Conditions and Number of People using Beach 

Although we did not assess beach conditions or number of people using the beach quantitatively, we observed a qualitative correlation between the number of people using the beach and the overall beach conditions.  The highest number of beach users was observed the afternoon of June 7th, and the lowest number was observed during the afternoon of June 15th.  Very few people were present at 9am on any of the sampling days.  Perhaps not surprisingly, the beach conditions deteriorated as the number of people increased.  For example, more trash and dirty diapers were observed on June 7th, the day that we observed the largest number of people.   

Similarly, the increase in bacteria concentration between 9am and 2pm was highest on June 7th and lowest on June 15th.    Figure 2 shows the relationship between the percent difference in afternoon vs. morning bacterial level and the relative number of people present on the beach.  The R2 value of 0.96 indicates a strong direct correlation between these two variables. 

Figure 2

Limitations

This study was limited, due to laboratory availability, to weekend days when the beach population was high.   Sampling on a day with no or few swimmers would have allowed us to measure the diurnal variation without having the data confounded by population of swimmers.   There were other single effects that could not be evaluated due to the many variables that could only be observed.   The weather was fairly consistent on the sampling days; more variability in wind direction would have resulted in more variation in water currents.   Water current observations improved as we learned how to more critically evaluate the surface and wave patterns and better procedures and equipment could be used for follow up studies.

   In addition, the algae observations were made in a relatively small area.   A more detailed map of the algae distribution in the swimming area might have shown a correlation between the bacteria levels and algae concentrations.   

   Lastly, the length of the study was confined into two weeks in the beginning of June.  The results would have more credibility with more sampling days over a longer period of time.  Extending the time period of the study would encompass more varied weather conditions and different people populations. 

Model Revisited

Our original model predicted that fecal coliform counts would change over the course of the day due to three factors: UV radiation, algae mass, and nearshore water currents.  We expected to find a correlation between one or more of these factors and fecal coliform counts in our samples. 

We predicted that UV radiation would kill off bacteria near the water’s surface because of its ability to disinfect water.  If this effect were significant, we expected that our morning samples would have the highest counts.  Our results, however, did not show this effect.  Even though we had relatively sunny weather on all three sampling days, our results showed that fecal coliform levels in morning samples were either the same or lower than those in the afternoon.  These results suggest that UV radiation does not significantly impact fecal coliform counts over the course of the day as compared to other factors.  

Although we expected to see a correlation between fecal coliform level and algae mass, our data showed no correlation.  This may have resulted from the fact that our observations of algae were limited to a relatively small area near the sample collection site.  A more detailed study over a longer period of time and covering a larger section of the beach may yield more conclusive results regarding the impact of algae on bacterial counts.

Nearshore water currents were also observed and current at the sampling site was measured where possible.  The data suggests that surface water currents, as influenced by wind speed and direction, affect fecal coliform levels by either scouring the beach during outflow conditions or building up levels during inflow conditions.   While sample Location 3 tended to have the most algae, as predicted because of proximity to the pier and an inability for current to clear it out, it was not the location with the highest fecal coliform count during this sampling period. 

One factor not included in our original model, was the number of people on the beach and in the water.   Due to the warm weather during the sampling period, the beach was crowded during the afternoon sampling time, while it was nearly empty during the morning sampling time.   We estimated the relative levels of people qualitatively for each sampling day.  The difference in fecal coliform levels between the a.m. and p.m. correlated most closely with the qualitative estimate of the number of people at the beach.  It is not clear from our data why this might this might occur.  We speculate that the activity of people near and in the water brings in contaminated sand and stirs up sediments from the bottom of the swimming area.  Figure 3 shows our revised model to include these two new related factors.  Although, as suggested above, our data provides some evidence to support this revised model, we recommend that these factors be further studied to determine causal effects.  

Figure 3:  Revised Model

Implications

Monroe County personnel take water samples for bacteria testing in the morning.   This does not appear to reflect the actual bacteria conditions during peak swimming time.  There are at least two possible mechanisms related to swimmers that could cause an increase in the bacteria levels in the afternoon:   (a) the swimmers stir up bottom sediment, which may contain bacteria and (b) swimmers are tracking the sand from the beach into the water.  Team Flagella has found that there are high bacteria levels in the sand.  If increased bacteria levels are due to contaminated sand, the impact on water quality of run-off from the proposed beach spray park should be evaluated prior to construction.

Our study shows that local, surface water currents as affected by wind direction can impact bacteria levels at various locations in the nearshore area.   Wind direction and speed could be added to the predictive model to improve accuracy although this is already included indirectly by evaluating the direction of the river plume.

     The Sandbaggers took some time to reflect on the nature of science as we finished our data collection experience this week.  We came to the conclusion that science is multi-faceted and cannot be described in a few words; instead, many descriptions apply.  Our thoughts on science, and a few stories of our experiences follow…

S -  surprising, statistically-intense at times, superb!

C - creative, collaborative, curious 

I - iterative, intriguing, inter-connected

E - engaging, exploring, explaining

N - natural, never neutral

C - controversial, challenging, changing

E - evaluating, expanding, evolving

     Before collecting data, we collaborated to decide what question(s) to explore through our research.  There was no shortage of ideas!  Our first list of research question options was very long because we were curious about so many aspects relating to water quality at the beach.  After listening to Charles Knauf’s summary of past research, and discussing options while drinking many pots of Jasmine Tea at the California Rollin’ Restaurant,  we reached a consensus to investigate:

• How do fecal coliform concentrations vary during the course of the day, and
• Can this variation be correlated with other variables, specifically UV index and water current speed & direction?

     We had several discussions on the nature and language of modeling relative to hypotheses, theories, assumptions, and other words commonly used to discuss perceptions of scientific concepts.  Our model of the many inter-related factors affecting water quality at Lake Ontario beach took the form of a concept map.  The map was a great tool for focusing and clarifying the mental models of each of our group members.  Through collaborative, argumentative (and fun!) discussions, a collective model was built.   Four heads were better than one in terms of building a sound, defensible concept map of the many variables affecting water quality.

     Our data collection adventures began on Saturday, June 7th.  The day dawned warm and windy - with plenty of biting flies to keep us on our toes.  We worked in teams of 2 to complete each of our total of 6 sampling events.  The transfer of equipment and samples, as well as the juggling of schedules for 4 busy women  required a high level of organization and cooperation,  and we did it!  Our last data samples were collected on the afternoon of June 15th.

     We discovered that science is iterative.  For example, our initial data collection worksheet was not bad for a first shot, but we found ourselves improving it almost immediately based on sampling experience.  The improvements were not major changes, but minor modifications to simplify the sequence of data collection and encourage consistency between sampling teams.  Everyone contributed to the process, and the final data collection sheet made the sampling process very efficient.  The first sampling event required 3 hours; the last one was completed in less than 1 hour.  

     An obstacle that we worked through successfully was finding representative weather data.  Our original plan was to use data collected by the National Weather Service at the Rochester Airport.  We soon realized (after a few very cold class sessions at the beach) that weather conditions near Lake Ontario can be very different from those at the airport.  A few phone calls to sailing friends yielded a “Eureka moment” when we learned that there was a weather station located right at the end of the Ontario Beach pier! 

     We were not quite as successful with a second obstacle involving Dissolved Oxygen probes.  The manufacturer’s directions for probe calibration involved collecting water samples in a container and shaking the container to fully saturate the water with oxygen as a first step.  This was used as the “100% saturation” calibration point.  We followed the procedure, but were surprised to find that the dissolved oxygen reading in the lake water was higher than the fully saturated sample in the container.  We reasoned that the temperature of the water in the container was probably rising as we exposed it to warm air.  Because oxygen is more soluble in cold water than in warm water,  we believe it is reasonable to assume that the higher readings in the lake water were due to lower temperatures in the lake vs. the container.  This theory remains untested.  Because dissolved oxygen was not central to our research question, we leave the validation of this theory and improvements in the calibration procedure to future investigators as a natural follow-up to our work. 

     The reaction from people at the beach during our sampling was a bit surprising.  Nobody commented on our fashionable hip-waders, but we fielded at least 50 questions from curious swimmers and beach combers who were intrigued by what we were doing (and maybe a little concerned about the quality of the water they were swimming in). They were relieved after we explained our mission and the mission of the camp to be held later in the summer.  In contrast, there were other beach users who demonstrated very little concern for water quality by leaving trash in the water and along the shoreline.  The image below is one of three dirty diapers that we observed on the first day of sampling.

     Kathryn encountered an unexpected challenge by coming to the rescue of a bicyclist involved in an accident near the beach.  It was fortunate that our data collection bag contained a first aid kit, as she was able to get him patched up in no time.  Our remaining challenge, that we fully expected, is to analyze the large amount of data that we collected in a span of a week.  We will be graphing data to look for correlations between variables, as well as calculating statistics to determine the precision of our protocols.  Stay tuned for the course finale, which we expect to publish on June 26th……

     The data was gathered on three nonconsecutive days at approximately 9 a.m. and 2 p.m.  Monroe County takes a sample at 9 a.m. to determine beach closures.  The samples taken at 2 p.m. correlate to the time when the beach is most populated and is usually the peak time for swimming.  We collected samples on three different days to in an attempt to include variable weather conditions.  Although we could not control for this variable, we hoped to see differences in air temperature, water current patterns, and precipitation.  For easy and accurate data recording, the group created a data sheet.  The data sheet was designed so that we could easily record our measurements.  It included a diagram of the shore on which we could draw our observations of the current patterns.  The original data sheet was adjusted after the first sampling event to better fit our process of gathering information. 

 Sandbaggers Data Sheet

     The water samples were taken from three locations.  Location 1 is in the middle of the swimming area.  The landmark for Location 1 was the shower-head next to the stairs leading to the beach on the west side of the bathhouse.  Location 1 is the area where the County takes their water sample for the bacterial counts.  Location 2 is on the west edge of the swimming area next to the area of the beach that is never open.  This location was defined by a light pole in front of the large gazebo.  Location 3 is next to the pier in a part of the beach that is never open due to high concentrations of algae and possible high bacteria levels.  This location was marked by a light pole between the two large trees next to the pier.  Location 1 was 1100 feet from the pier, Location 2 was 550 feet from the pier, and Location 3 was 67 feet from the pier. The sample locations were determined by landmarks for consistency during each sampling event. 

     The date and time along with general weather conditions, such as percent of cloud cover and wind direction, were recorded before the samples were taken.  The percentage of cloud cover was estimated by a mutual agreement between group members, and compared to readings posted on http://www.intellicast.com/Local/Weather.aspx?location=USNY1232 for validation.  Pictures were taken of the beach and water, and observations were recorded about the river plume and noticable odors.   

     The first samples were taken from Location 1 moving eastward down the beach to Location 3.  This was the best way to ensure that bacteria was not tracked from a closed area to a swimming area.  The algae level was qualitatively described based on visual observations and documented by pictures in three different areas from the shore to the sample site: 0-3 feet from shore, 3-6 feet from shore, and the sample site.  Once at the sample site, water samples were taken for bacteria testing.  Samples were kept in a cooler packed with ice to prevent more bacteria growth.    

     At each sample location the water depth was 29 inches; the sample was taken six inches below the surface at knee level.  This was a fairly precise depth because the members of the group were about the same height.  Care was taken not to disturb bottom sediment while walking out to the sampling site.  We choose the six inch depth to obtain a representative determination of the overall bacteria count of the swimming water and not just the surface water.

    There were three water samples gathered at each location to measure sampling induced variability at each location.  To gather the water sample, we used whirl-pack bags - easy and minimally contaminated by the group member.  The turbidity was measure using the Secchi Disk.  The turbidiy and cloud cover, along with the predicted UV index posted at http://www.intellicast.com/Local/Weather.aspx?location=USNY1232 was used to estimate the total UV exposure.  

     Observations of the water current direction and velocity were recorded at the sample site.   The water current was recorded from visual observations of the waves and a floating orange-colored ball with attached string.  The direction of the current based on the direction the orange ball float was determined and agreed upon by sampling team.  The orange ball was timed for its movement of six feet at each of the locations.  Once the ball hit the water, the stopwatch was started then when the string attached to the ball was tight the stop watched was stopped.  The string was marked and held at the six feet mark for an accurate measure.  A video made of the orange ball to demonstrate the process of timing the ball.  The direction of the waves was drawn on the data sheet to show the current direction at far shore, nearshore, and waves onto shore were recorded.   Wind direction and speed at sampling time were recorded from the NOAA weather station at the end of the pier on the east side of the river, see previous post First Day of Data Collection.

     Once the location specific measurements were taken, additional measurements were gathered at Location 1 at the location of the water sample.  The information was collected using the Pasco Xplorer and each measurement had a unique probe.  The temperature of the air and water were measured using the Pasco Xplorer and recorded on the data sheet.  The pH and conductivity were also measured using the Pasco Xplorer.  The last data collected from Location 1 was the dissolved oxygen.  The maximum saturation of the water is needed to compare to the actual dissolved oxygen of the water.  This was gathered by collecting a water sample in a bottle and shaking the bottle.  This saturates the water sample, then checks the calibration of the dissolved oxygen probe.  The number was recorded on the data sheet.  Then we put the dissolved oxygen probe in the water and this number was recorded.  This data was then compared to a chart of what the dissolved oxygen number should be based on the air temperature and the barometric pressure.

     Once these measurements were completed, any significant final observations were recorded.  The water samples were then brought to the laboratory to be plated.  The plated water samples were to test for the amount of fecal coliform in the water.  An aliquot size of 2.4-2.6 mL was obtained from each sample bag using a sterile pipette.  The aliquot was then poured into a bottle of Coliscan Easygel and, after mixing, poured into a culturing plate.  The plates were then incubated at 37 deg. C. and for 24 hours.  The plates were kept refrigerated until bacterial counts were evaluated.  Both total coliform and fecal coliform counts were recorded.     

     Items we used each sampling event were 9 whirl-pack bags, 3 large zip lock bags, Pasco Xplorer, Secchi Disk, and a orange ball with string.  We needed to bring with us the barometric pressure and UV index readings, camera, cooler, data sheet, hip waders and stopwatch. 

     We were up at the beach bright and early this morning.    We also saw Team Flagella there!   After freezing for the past two weeks during class, it was wonderful that the temperature was above 80 even at 8:30 in the morning.   We could have done without the nasty, ankle-biting flies though.  It was very windy with gusts that blew sand and some of our papers around.  The minute we stepped off the beach and into the water, it was cooler as the air moved over the 13 deg Celsius (55 deg Farenheit) water.  

     The thunder and lightning kept us out of the water on Thursday evening which meant that we had to do more protocol development today than we had anticipated.   We think we have a good sampling plan and are looking forward to seeing our results.   We hope to have different weather conditions but tomorrow looks like it will be more of the same.

     We did a bit of exploring and walked out the pier to see the river plume.   I also took a picture of our new favorite weather station which provides great data, current and historical, from the area where we are sampling.

You can access this data at http://www.ndbc.noaa.gov/station_page.php?station=rprn6

       Currently, samples used to evaluate water quality at Ontario Beach are taken during the morning hours, taken at approximately 9 a.m.  If E. coli levels change significantly during the course of the day, these morning samples may not accurately reflect water quality during the peak swimming hours, which tend to occur in the afternoon.  In addition, once the beach is closed, it cannot reopen until the E. coli levels fall below 235 mpn/100 mL  (Knaupf, presentation, 2008), this again based only on morning sampling.  Our study is designed to address this gap in the current research by posing the following question:

How does water quality as indicated by bacterial count vary over the course of the day and can this variation be correlated with measurable conditions?

        Figure 2 shows our model of the factors we believe impact water quality as indicated by bacterial count.  Our model suggests that, for a given sampling site, bacterial counts are affected by the time of day at which the sample was collected. 

 Figure 2

        In this model the bacteria count is the dependent variable and is influenced by three major factors: UV exposure, nearshore water current, and algae mass.  As suggested by Figure 2, the time of day directly impacts the UV exposure and the nearshore water current.  These two factors in turn affect localized algae concentrations. 

        The UV exposure is a term used here to describe amount of solar radiation penetrating through the water.  We will qualitatively evaluate UV exposure based on UV index, a qualitative assessment of cloud cover and measured turbidity.  We will also be experimenting with Solar Print Paper in the hope of developing a more qualitative test of UV conditions as the time of sampling.  Our hypothesis is that UV exposure has a negative impact on bacterial counts (i.e. as UV exposure increases bacterial counts will decease).  

        The nearshore water current impacts both the bacterial count and algae mass.  In general with water flow away from the beach, there is a greater chance for dilution of both algae and bacterial concentrations.   At Ontario Beach the pier creates a significant barrier to flow.  We expect that our results will be impacted by the effect of this pier.  We also expect that the water surface will be calmer in the morning than in the afternoon.  For this study, the nearshore water current will be evaluated based on our measurements of surface current speed and direction and locally available data on wind speed and direction.  Our hypothesis is that bacterial count will increase if current is flowing from the northwest because algae will be pushed into the swimming area and the pier will prevent it from flowing out.  

        Algae provide both nutrients and a medium for bacterial growth.  We will measure this factor qualitatively based on our observations of the water at each sampling site.  Our hypothesis is that bacterial counts will tend to be higher in the areas with higher algae levels.

       Additional measurements of water temperature, dissolved oxygen, conductivity and pH will be made; however, we do not expect these conditions to greatly impact the model over the time and locations we will be sampling.  General observations will be made of odor, organic debris, wildlife, and river plume direction to determine possible impacts on results. 

Summary of Investigation Design 

Independent Variables

- time of day:  9 a.m. and 2 p.m.

- sample site locations: three locations along beach

Measured Conditions

- UV index

- turbidity

- current speed and direction

- wind speed and direction

Factors

- UV exposure: qualitative based UV index, local cloud cover and turbidity

- nearshore water current: observed using floating balls

- algae mass: qualitative visual observation

Dependent Variable

- bacterial count: three water samples will be collected at each sampling site

Additional Measurements and Observations

- water temperature

- dissolved oxygen

- conductivity

- pH

- general observations: odors, beach debris and river plume 

Birds at Ontario Beach  - Source:   mmahaffie - flickr.com

     In 1967, the three Lake Ontario beaches located in Monroe County were closed by the New York State Health Department because samples taken in 1966 showed high levels of fecal coliform.  The beaches were closed until 1976 when a model to predict water quality at Ontario Beach was developed and approved.   (Knaupf, April 2008)   These beaches are unique in that they are located within the Rochester embayment formed by an indent in the southern coastline of Lake Ontario (see Figure 1 below);  this area is the drainage basin for more than 3,000 square miles including the Genesee River basin.   (United States Environmental Protection Agency [EPA], 2008)    Ontario Beach, located just west of the mouth of the Genesee River, will be the focus of our investigation.

Figure 1 - Source: EPA Website

     The shoreline profile which forms the Rochester Embayment provides a unique ecosystem sheltered from the currents of the Great Lakes Seaway.     The west end of the Rochester Embayment diverts the general current away from this drainage basin reducing the rate of flow out of the area.  The water flow is from the western end to the St. Laurence River across Lake Ontario but the surface currents that have more impact on the coastal areas are influenced primarily by wind direction.   (Yerubandi et al., 2008)

     The Genesee river is the major tributary to this Embayment.   The river has higher bacterial counts than the lake but current levels are significantly lower than historical numbers due to sewage mitigation and control of stormwater overflow events using the deep tunnel system in the bedrock under Rochester.    The high solids content of the river, mostly from soil erosion, adds to the turbidity of the lake and a plume of brown river water can be seen extending from the end of the pier into Lake Ontario.   The direction of this plume is used to determine the impact of the river on beach water quality.    (Knaupf, presentation, May 2008) 

     At a recent meeting of the Rochester Committee for Scientific Information, Charles Knaupf presented the predictive models used for the 2007 season at Ontario Beach and Durand Beach, reasons for beach closings, and summarized the accuracy of the predictive model.   (Knaupf, April 2008)  The 18-24 hours needed to culture a bacterial sample mandate the use of a model based on measurable factors that predict bacteria level.    There are five major factors for both models:    Genesee River flow and plume direction, rainfall over the center of Rochester and the immediate beach area, quantity of algae or other organic debris, water clarity, and prior day bacterial level test result.    Criteria for Ontario Beach are presented in Table 1 below. 

 Table 1 (Summarized from Knaupf, April 2008)

Ontario Beach:   2007 Operating Criteria

     In 2007, Ontario Beach was open for swimming in all sections of the beach for 19 days (25% of the season), some sections were open for 30 days (40%), and all sections were closed for 26 days (35%).    For the 2007 season, the model did not accurately predict the beach conditions on 31 days (41%);  fifteen days open with a dirty beach and 16 days closed with a clean beach.    The primary reasons for closure were algae or other organic debris (54%), poor water clarity (22%), and prior day bacteria levels (19%). 

      Algae is highly correlated with bacteria count at this beach and 1,500 cubic feet of algae in any section of the beach results in closure of that section.   Three types of algae are predominant in this ecosystem, Cladophora, Ulothrix, and Spirogyra;   Cladophora is the major factor during the swimming season.   (Knaupf, presentation, May 2008)    One study (Byappanahalli, Shively, Nevers, Sadowsky, & Whitman, 2003) indicated that Cladophora algae can serve as both a growth medium and a nutrient source for E. Coli and enterococci populations.    The presence of these bacteria is associated with the possibility of other human pathogens in the water.   (Ishii et al., 2006)   The half-mile long pier at the east end of the beach traps algae and prevents it from being swept out to the lake.   

     Rainfall was not a primary factor in beach closings for 2007 because it was a hot, dry year.   It is included in the predictive model because of the wash off of nearshore contamination.     (Knaupf, presentation to class,  5/29/08)   A report on sources of fecal pollution at Hamilton Harbour at the western end of Lake Ontario (Edge & Hill, 2007) demonstrates the contribution of bird droppings to fecal coliform level in the water off that beach.

     When asked about model inaccuracy of 41% and gaps in understanding of the ecosystem, Charles Knaupf identified the impact of birds and other wild life on bacterial level and variability in the data sampling as factors that are not well understood.    Although fecal coliform bacteria are an indicator of the presence of human pathogens, the bacteria may originate from birds and other wildlife as well as human sources.  The origins of the indicator bacteria present at Ontario Beach have not been studied in depth to date.    Bacterial sample results are generally higher in the morning and decrease later in the day.   (Knaupf, presentation, 2008)  Potential causes of this variation are exposure to UV light and other factors related to sample location and conditions.

References

 Byappanahalli, M. N., Shively, D. A., Nevers, M. B., Sadowsky, M. J., & Whitman, R. L. (2003). Growth and survival of Escherichia coli and enterococci populations in the macro-alga Cladophora (Chlorophyta). FEMS Microbiology Ecology, 46(2), 203-211.

Edge, T. A., & Hill, S. (2007). Multiple lines of evidence to identify the sources of fecal pollution at a freshwater beach in Hamilton Harbour, Lake Ontario. Water Research, 41(16), 3585-3594.

Ishii, S., Yan, T., Shively, D. A., Byappanahalli, M. N., Whitman, R. L., & Sadowsky, M. J. (2006). Cladophora (Chlorophyta) spp. Harbor Human Bacterial Pathogens in Nearshore Water of Lake Michigan. Applied and Environmental Microbiology, 72(7), 4545.

Knaupf, C. (2008, April). Predictive Models for Beach Operation in Monroe County:  Historical Perspectives and Recent Data presented at the meeting of the Rochester Committee for Scientific Information.   Retrieved May 24, 2008, from http://www.rcsiweb.org/bulletins/monroe_county_beaches_knaupf_may2008.pdf

 United States Environmental Protection Agency.   (n.d.)   Rochester Embayment Area of Concern.    Retrieved on May 24, 2008 from http://www.epa.gov/glnpo/aoc/rochester.html

Yerubandi, R. R., Lam, D., Schertzer, W., Leon, L., Zhao, J., & Thompson, A. (2008, May).    A modelling framework for environmental prediction in the Great Lakes.    Paper presented at the 2008 Congress of the Canadian Meteorological and Oceanographic Society,  Kelowa, British Columbia.

Great Lakes Commission Website

This website has general overview of Lake Ontario with two main concerns.  The first is drinking water and the effects of microbial and chemical contamination.  The second covers recreation and the reasons for beach closings.  There are also links at the bottom for fish consumption and additional information on New York and Lake Ontario.  

For perspective, Ontario Beach has been cited by the EPA as one of several ’Loss of Beneficial Use’ concerns in the overall Rochester Embayment area.  Corrective actions, as well as some good images, can be found at this site:  http://www.epa.gov/glnpo/aoc/rochester.html

The picture is of Cladophora Glomerata.    In May, 2002,  experts on the Great Lake basin with specific expertise on Lake Ontario convened for a Lake Ontario Algae Cause and Solution Workshop.   These proceedings contain both the basic science and a historical narrative about causes and control of algae in Lake Ontario.

Side note:  In rereading the syllabus for the course, I see that these blogs are not intended to be discussion or collaborative blogs but rather reporting blogs.   I’ll have to talk to our instructor.   This post may be out of bounds for several reasons.

We have a list of topics to research and a Question and Model blog entry due on 5/31.   I’ve found many articles in local newspapers but by far the best information was posted by Chris on the Team Flagella blog.    It’s a detailed presentation given last month on Predictive Models for Beach Operation by Charles Knaupf.    Check out the link at the top of their page but if that doesn’t work, you could also try clicking here .  

Thanks Chris for letting us know about it via comment on this blog.   I hope the trackback works!